Class: Moxml::Adapter::Rexml

Inherits:
Base
  • Object
show all
Defined in:
lib/moxml/adapter/rexml.rb

Class Method Summary collapse

Methods inherited from Base

create_cdata, create_comment, create_declaration, create_doctype, create_element, create_namespace, create_processing_instruction, create_text, patch_node

Methods included from XmlUtils

#encode_entities, #normalize_xml_value, #validate_comment_content, #validate_declaration_encoding, #validate_declaration_standalone, #validate_declaration_version, #validate_element_name, #validate_pi_target, #validate_prefix, #validate_uri

Class Method Details

.add_child(element, child) ⇒ Object 



229
230
231
232
233
234
235
236
 
# File 'lib/moxml/adapter/rexml.rb', line 229

def add_child(element, child)
  case child
  when String
    element.add_text(child)
  else
    element.add(child)
  end
end

.add_next_sibling(node, sibling) ⇒ Object 



248
249
250
251
 
# File 'lib/moxml/adapter/rexml.rb', line 248

def add_next_sibling(node, sibling)
  parent = node.parent
  parent.insert_after(node, sibling)
end

.add_previous_sibling(node, sibling) ⇒ Object 



238
239
240
241
242
243
244
245
246
 
# File 'lib/moxml/adapter/rexml.rb', line 238

def add_previous_sibling(node, sibling)
  parent = node.parent
  # caveat: Rexml fails if children belong to the same parent and are already in a correct order
  # example: "<root><a/><b/></root>"
  # add_previous_sibling(node_b, node_a)
  # result: "<root><b/><a/></root>"
  # expected result: "<root><a/><b/></root>"
  parent.insert_before(node, sibling)
end

.at_xpath(node, expression, namespaces = {}) ⇒ Object 



411
412
413
414
 
# File 'lib/moxml/adapter/rexml.rb', line 411

def at_xpath(node, expression, namespaces = {})
  results = xpath(node, expression, namespaces)
  results.first
end

.attribute_element(attribute) ⇒ Object 



195
196
197
 
# File 'lib/moxml/adapter/rexml.rb', line 195

def attribute_element(attribute)
  attribute.element
end

.attributes(element) ⇒ Object 



187
188
189
190
191
192
193
 
# File 'lib/moxml/adapter/rexml.rb', line 187

def attributes(element)
  return [] unless element.respond_to?(:attributes)

  # Only return non-namespace attributes
  element.attributes.values
         .reject { |attr| attr.prefix.to_s.start_with?("xmlns") }
end

.cdata_content(node) ⇒ Object 



296
297
298
 
# File 'lib/moxml/adapter/rexml.rb', line 296

def cdata_content(node)
  node.value
end

.children(node) ⇒ Object 



113
114
115
116
117
118
119
120
121
122
123
124
125
 
# File 'lib/moxml/adapter/rexml.rb', line 113

def children(node)
  return [] unless node.respond_to?(:children)

  # Get all children and filter out empty text nodes between elements
  result = node.children.reject do |child|
    child.is_a?(::REXML::Text) &&
      child.to_s.strip.empty? &&
      !(child.next_sibling.nil? && child.previous_sibling.nil?)
  end

  # Ensure uniqueness by object_id to prevent duplicates
  result.uniq(&:object_id)
end

.comment_content(node) ⇒ Object 



288
289
290
 
# File 'lib/moxml/adapter/rexml.rb', line 288

def comment_content(node)
  node.string
end

.create_document(_native_doc = nil) ⇒ Object 



24
25
26
 
# File 'lib/moxml/adapter/rexml.rb', line 24

def create_document(_native_doc = nil)
  ::REXML::Document.new
end

.create_native_cdata(content) ⇒ Object 



36
37
38
 
# File 'lib/moxml/adapter/rexml.rb', line 36

def create_native_cdata(content)
  ::REXML::CData.new(content.to_s)
end

.create_native_comment(content) ⇒ Object 



40
41
42
 
# File 'lib/moxml/adapter/rexml.rb', line 40

def create_native_comment(content)
  ::REXML::Comment.new(content.to_s)
end

.create_native_declaration(version, encoding, standalone) ⇒ Object 



49
50
51
 
# File 'lib/moxml/adapter/rexml.rb', line 49

def create_native_declaration(version, encoding, standalone)
  ::REXML::XMLDecl.new(version, encoding&.downcase, standalone)
end

.create_native_doctype(name, external_id, system_id) ⇒ Object 



53
54
55
56
57
58
59
60
61
62
63
64
65
 
# File 'lib/moxml/adapter/rexml.rb', line 53

def create_native_doctype(name, external_id, system_id)
  return nil unless name

  parts = [name]
  if external_id
    parts.concat(["PUBLIC", %("#{external_id}")])
    parts << %("#{system_id}") if system_id
  elsif system_id
    parts.concat(["SYSTEM", %("#{system_id}")])
  end

  ::REXML::DocType.new(parts.join(" "))
end

.create_native_element(name) ⇒ Object 



28
29
30
 
# File 'lib/moxml/adapter/rexml.rb', line 28

def create_native_element(name)
  ::REXML::Element.new(name.to_s)
end

.create_native_namespace(element, prefix, uri) ⇒ Object

add a namespace definition, keep the element name unchanged



348
349
350
351
 
# File 'lib/moxml/adapter/rexml.rb', line 348

def create_native_namespace(element, prefix, uri)
  element.add_namespace(prefix.to_s, uri)
  ::REXML::Attribute.new(prefix.to_s, uri, element)
end

.create_native_processing_instruction(target, content) ⇒ Object 



44
45
46
47
 
# File 'lib/moxml/adapter/rexml.rb', line 44

def create_native_processing_instruction(target, content)
  # Clone strings to avoid frozen string errors
  ::REXML::Instruction.new(target.to_s.dup, content.to_s.dup)
end

.create_native_text(content) ⇒ Object 



32
33
34
 
# File 'lib/moxml/adapter/rexml.rb', line 32

def create_native_text(content)
  ::REXML::Text.new(content.to_s, true, nil)
end

.declaration_attribute(node, name) ⇒ Object 



266
267
268
269
270
271
272
273
274
275
 
# File 'lib/moxml/adapter/rexml.rb', line 266

def declaration_attribute(node, name)
  case name
  when "version"
    node.version
  when "encoding"
    node.encoding
  when "standalone"
    node.standalone
  end
end

.document(node) ⇒ Object 



179
180
181
 
# File 'lib/moxml/adapter/rexml.rb', line 179

def document(node)
  node.document
end

.duplicate_node(node) ⇒ Object 



107
108
109
110
111
 
# File 'lib/moxml/adapter/rexml.rb', line 107

def duplicate_node(node)
  # Make a complete duplicate of the node
  # https://stackoverflow.com/questions/23878384/why-the-original-element-got-changed-when-i-modify-the-copy-created-by-dup-meth
  Marshal.load(Marshal.dump(node))
end

.get_attribute(element, name) ⇒ Object 



217
218
219
 
# File 'lib/moxml/adapter/rexml.rb', line 217

def get_attribute(element, name)
  element.attributes.get_attribute(name)
end

.get_attribute_value(element, name) ⇒ Object 



221
222
223
 
# File 'lib/moxml/adapter/rexml.rb', line 221

def get_attribute_value(element, name)
  element.attributes[name]
end

.inner_text(node) ⇒ Object 



327
328
329
330
331
332
333
 
# File 'lib/moxml/adapter/rexml.rb', line 327

def inner_text(node)
  # Get direct text children only, filter duplicates
  text_children = node.children
                      .select { _1.is_a?(::REXML::Text) }
                      .uniq(&:object_id)
  text_children.map(&:value).join
end

.namespace(node) ⇒ Object 



370
371
372
373
374
375
376
377
 
# File 'lib/moxml/adapter/rexml.rb', line 370

def namespace(node)
  prefix = node.prefix
  uri = node.namespace(prefix)
  return if prefix.to_s.empty? && uri.to_s.empty?

  owner = node.is_a?(::REXML::Attribute) ? node.element : node
  ::REXML::Attribute.new(prefix, uri, owner)
end

.namespace_definitions(node) ⇒ Object 



379
380
381
382
383
 
# File 'lib/moxml/adapter/rexml.rb', line 379

def namespace_definitions(node)
  node.namespaces.map do |prefix, uri|
    ::REXML::Attribute.new(prefix.to_s, uri, node)
  end
end

.namespace_prefix(node) ⇒ Object 



362
363
364
 
# File 'lib/moxml/adapter/rexml.rb', line 362

def namespace_prefix(node)
  node.name unless node.name == "xmlns"
end

.namespace_uri(node) ⇒ Object 



366
367
368
 
# File 'lib/moxml/adapter/rexml.rb', line 366

def namespace_uri(node)
  node.value
end

.next_sibling(node) ⇒ Object 



131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
 
# File 'lib/moxml/adapter/rexml.rb', line 131

def next_sibling(node)
  current = node.next_sibling

  # Skip empty text nodes and duplicates
  seen = Set.new
  while current
    if current.is_a?(::REXML::Text) && current.to_s.strip.empty?
      current = current.next_sibling
      next
    end

    # Check for duplicates
    if seen.include?(current.object_id)
      current = current.next_sibling
      next
    end

    seen.add(current.object_id)
    break
  end

  current
end

.node_name(node) ⇒ Object 



96
97
98
99
100
101
102
103
104
105
 
# File 'lib/moxml/adapter/rexml.rb', line 96

def node_name(node)
  case node
  when ::REXML::Element, ::REXML::DocType
    node.name
  when ::REXML::XMLDecl
    "xml"
  when ::REXML::Instruction
    node.target
  end
end

.node_type(node) ⇒ Object 



71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
 
# File 'lib/moxml/adapter/rexml.rb', line 71

def node_type(node)
  case node
  when ::REXML::Document then :document
  when ::REXML::Element then :element
  when ::REXML::CData then :cdata
  when ::REXML::Text then :text
  when ::REXML::Comment then :comment
  when ::REXML::Attribute then :attribute # but in fact it may be a namespace as well
  when ::REXML::Namespace then :namespace # we don't use this one
  when ::REXML::Instruction then :processing_instruction
  when ::REXML::DocType then :doctype
  when ::REXML::XMLDecl then :declaration
  else :unknown
  end
end

.parent(node) ⇒ Object 



127
128
129
 
# File 'lib/moxml/adapter/rexml.rb', line 127

def parent(node)
  node.parent
end

.parse(xml, options = {}) ⇒ Object 



13
14
15
16
17
18
19
20
21
22
 
# File 'lib/moxml/adapter/rexml.rb', line 13

def parse(xml, options = {})
  native_doc = begin
    ::REXML::Document.new(xml)
  rescue ::REXML::ParseException => e
    raise Moxml::ParseError.new(e.message, line: e.line) if options[:strict]

    create_document
  end
  DocumentBuilder.new(Context.new(:rexml)).build(native_doc)
end

.prepare_xpath_namespaces(node) ⇒ Object

not used at the moment but may be useful when the xpath is upgraded to work with namespaces



387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
 
# File 'lib/moxml/adapter/rexml.rb', line 387

def prepare_xpath_namespaces(node)
  ns = {}

  # Get all namespace definitions in scope
  all_ns = namespace_definitions(node)

  # Convert to XPath-friendly format
  all_ns.each do |prefix, uri|
    if prefix.to_s.empty?
      ns["xmlns"] = uri
    else
      ns[prefix] = uri
    end
  end

  ns
end

.previous_sibling(node) ⇒ Object 



155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
 
# File 'lib/moxml/adapter/rexml.rb', line 155

def previous_sibling(node)
  current = node.previous_sibling

  # Skip empty text nodes and duplicates
  seen = Set.new
  while current
    if current.is_a?(::REXML::Text) && current.to_s.strip.empty?
      current = current.previous_sibling
      next
    end

    # Check for duplicates
    if seen.include?(current.object_id)
      current = current.previous_sibling
      next
    end

    seen.add(current.object_id)
    break
  end

  current
end

.processing_instruction_content(node) ⇒ Object 



308
309
310
 
# File 'lib/moxml/adapter/rexml.rb', line 308

def processing_instruction_content(node)
  node.content
end

.processing_instruction_target(node) ⇒ Object 



304
305
306
 
# File 'lib/moxml/adapter/rexml.rb', line 304

def processing_instruction_target(node)
  node.target
end

.remove(node) ⇒ Object 



253
254
255
 
# File 'lib/moxml/adapter/rexml.rb', line 253

def remove(node)
  node.remove
end

.remove_attribute(element, name) ⇒ Object 



225
226
227
 
# File 'lib/moxml/adapter/rexml.rb', line 225

def remove_attribute(element, name)
  element.delete_attribute(name.to_s)
end

.replace(node, new_node) ⇒ Object 



257
258
259
 
# File 'lib/moxml/adapter/rexml.rb', line 257

def replace(node, new_node)
  node.replace_with(new_node)
end

.replace_children(element, children) ⇒ Object 



261
262
263
264
 
# File 'lib/moxml/adapter/rexml.rb', line 261

def replace_children(element, children)
  element.children.each(&:remove)
  children.each { |child| element.add(child) }
end

.root(document) ⇒ Object 



183
184
185
 
# File 'lib/moxml/adapter/rexml.rb', line 183

def root(document)
  document.root
end

.serialize(node, options = {}) ⇒ Object 



416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
 
# File 'lib/moxml/adapter/rexml.rb', line 416

def serialize(node, options = {})
  output = String.new

  if node.is_a?(::REXML::Document)
    # Always include XML declaration
    decl = node.xml_decl || ::REXML::XMLDecl.new("1.0", options[:encoding] || "UTF-8")
    decl.encoding = options[:encoding] if options[:encoding]
    output << "<?xml"
    output << %( version="#{decl.version}") if decl.version
    output << %( encoding="#{decl.encoding}") if decl.encoding
    output << %( standalone="#{decl.standalone}") if decl.standalone
    output << "?>"
    # output << "\n"

    if node.doctype
      node.doctype.write(output)
      # output << "\n"
    end

    # Write processing instructions
    node.children.each do |child|
      next unless [::REXML::Instruction, ::REXML::CData, ::REXML::Comment, ::REXML::Text].include?(child.class)

      write_with_formatter(child, output, options[:indent] || 2)
      # output << "\n"
    end

    write_with_formatter(node.root, output, options[:indent] || 2) if node.root
  else
    write_with_formatter(node, output, options[:indent] || 2)
  end

  output.strip
end

.set_attribute(element, name, value) ⇒ Object 



199
200
201
202
 
# File 'lib/moxml/adapter/rexml.rb', line 199

def set_attribute(element, name, value)
  element.attributes[name&.to_s] = value
  ::REXML::Attribute.new(name&.to_s, value.to_s, element)
end

.set_attribute_name(attribute, name) ⇒ Object 



204
205
206
207
208
209
210
211
 
# File 'lib/moxml/adapter/rexml.rb', line 204

def set_attribute_name(attribute, name)
  old_name = attribute.expanded_name
  attribute.name = name
  # Rexml doesn't change the keys of the attributes hash
  element = attribute.element
  element.attributes.delete(old_name)
  element.attributes << attribute
end

.set_attribute_value(attribute, value) ⇒ Object 



213
214
215
 
# File 'lib/moxml/adapter/rexml.rb', line 213

def set_attribute_value(attribute, value)
  attribute.normalized = value
end

.set_cdata_content(node, content) ⇒ Object 



300
301
302
 
# File 'lib/moxml/adapter/rexml.rb', line 300

def set_cdata_content(node, content)
  node.value = content.to_s
end

.set_comment_content(node, content) ⇒ Object 



292
293
294
 
# File 'lib/moxml/adapter/rexml.rb', line 292

def set_comment_content(node, content)
  node.string = content.to_s
end

.set_declaration_attribute(node, name, value) ⇒ Object 



277
278
279
280
281
282
283
284
285
286
 
# File 'lib/moxml/adapter/rexml.rb', line 277

def set_declaration_attribute(node, name, value)
  case name
  when "version"
    node.version = value
  when "encoding"
    node.encoding = value
  when "standalone"
    node.standalone = value
  end
end

.set_namespace(element, ns) ⇒ Object

add a namespace prefix to the element name AND a namespace definition



354
355
356
357
358
359
360
 
# File 'lib/moxml/adapter/rexml.rb', line 354

def set_namespace(element, ns)
  prefix = ns.name.to_s.empty? ? "xmlns" : ns.name.to_s
  element.add_namespace(prefix, ns.value) if element.respond_to?(:add_namespace)
  element.name = "#{prefix}:#{element.name}"
  owner = element.is_a?(::REXML::Attribute) ? element.element : element
  ::REXML::Attribute.new(prefix, ns.value, owner)
end

.set_node_name(node, name) ⇒ Object 



87
88
89
90
91
92
93
94
 
# File 'lib/moxml/adapter/rexml.rb', line 87

def set_node_name(node, name)
  case node
  when ::REXML::Element
    node.name = name.to_s
  when ::REXML::Instruction
    node.target = name.to_s
  end
end

.set_processing_instruction_content(node, content) ⇒ Object 



312
313
314
 
# File 'lib/moxml/adapter/rexml.rb', line 312

def set_processing_instruction_content(node, content)
  node.content = content.to_s
end

.set_root(doc, element) ⇒ Object 



67
68
69
 
# File 'lib/moxml/adapter/rexml.rb', line 67

def set_root(doc, element)
  doc.add_element(element)
end

.set_text_content(node, content) ⇒ Object 



335
336
337
338
339
340
341
342
343
344
345
 
# File 'lib/moxml/adapter/rexml.rb', line 335

def set_text_content(node, content)
  case node
  when ::REXML::Text, ::REXML::CData
    node.value = content.to_s
  when ::REXML::Element
    # Remove existing text nodes to prevent duplicates
    node.texts.each(&:remove)
    # Add new text content
    node.add_text(content.to_s)
  end
end

.text_content(node) ⇒ Object 



316
317
318
319
320
321
322
323
324
325
 
# File 'lib/moxml/adapter/rexml.rb', line 316

def text_content(node)
  case node
  when ::REXML::Text, ::REXML::CData
    node.value.to_s
  when ::REXML::Element
    # Get all text nodes, filter out duplicates, and join
    text_nodes = node.texts.uniq(&:object_id)
    text_nodes.map(&:value).join
  end
end

.xpath(node, expression, _namespaces = {}) ⇒ Object 



405
406
407
408
409
 
# File 'lib/moxml/adapter/rexml.rb', line 405

def xpath(node, expression, _namespaces = {})
  node.get_elements(expression).to_a
rescue ::REXML::ParseException => e
  raise Moxml::XPathError, e.message
end